Imagine that you are taking a picture of a product you want to promote on a brochure or on a website for advertising. From which angle should a picture be taken to give the best impression of the three-dimensional object?
Previous studies on the relationship between the object angle and goodness of view have shown that an oblique view (or a three-quarter view) has more advantages over a straight frontal or lateral view. Blanz et al. [1
] used interactive three-dimensional computer graphics and asked participants to choose the viewpoint from which they would take a photograph of common objects for a brochure to give the best possible impression (Experiment 1) or to choose the viewpoint from which they spontaneously formed mental images of the objects (Experiment 2). The results showed that most of the participants (all German residents) preferred oblique views to the straight front or side views, particularly in the photography task. Niimi and Yokosawa [2
] asked 10 Japanese participants to rate the view goodness of 18 common symmetrical objects rotated by 10 different angles from the straight front; they found that the three-quarter views, such as rotated by 27° and 45°, were rated better than other views. Nonose et al. [3
] further reported that although oblique views were generally rated better than the front and side views, different objects had different best viewing angles, and the objects could be classified into three clusters based on view goodness rating patterns across viewing angles: front-preferred, oblique-preferred, and side-preferred objects. Three-dimensionality ratings and recognition performance were the highest for oblique views for all clusters. Therefore, Nonose et al. suggested that familiarity, rather than recognition efficiency, was a major determinant of view goodness, although the most familiar views did not necessarily equal the best view.
Given that oblique views have advantages over other views, in which direction is an object to be rotated to obtain a better view? Intuitively, it does not matter which side of an object is presented to the viewer if the object is symmetric because both images contain exactly the same amount of information (i.e., mirror-images). However, research has repeatedly demonstrated that an object facing one direction is preferred over the same object facing the opposite direction. Most interestingly, the preferred direction varies considerably across cultures [4
] (see for a recent review). In particular, participants’ native reading and writing directions have attracted attention [5
]. Typically, in most studies, participants are presented with an object’s image and its mirror reversal and asked to decide which member looks better or is more aesthetically pleasing. In one of the earliest studies, Nachson et al. [7
] used 29 pairs of the left and right profiles of human faces and bodies, presented side by side, and asked 45 Arabic, 62 Hebrew, and 31 Russian readers to choose the member of a pair that looked more beautiful. The results showed that participants whose native language was Hebrew (a right-to-left script) preferred profiles that turned to the left, while participants whose native language was Russian (a left-to-right script) preferred profiles that turned to the right. Participants whose native language was Arabic (a right-to-left script) showed a nonsignificant leftward preference. These findings were regarded as evidence that aesthetic preference was affected by habitual reading direction. Chokron and De Agostini [8
] presented 30 mirror-image pairs of drawings located one above the other. They asked 81 French and 81 Israeli participants to indicate which stimulus was more aesthetically pleasing or interesting to look at. All the participants were right-handed. For static objects (10 pairs), French participants (left-to-right readers) preferred pictures facing to the right over its mirror-image, whereas Israeli participants, who read Hebrew from right to left, preferred pictures facing to the left over its counterpart. French readers’ rightward bias was replicated in a subsequent study [9
]. Ishii et al. [10
] conducted a similar survey with 50 Japanese and 50 Australian individuals. The results showed that Japanese people preferred left-facing objects, whereas Australian people preferred right-facing objects. It is noteworthy that the effect size (calculated by the current authors as the absolute deviation from no bias) was larger for the Japanese (d
= −0.82, from zero to leftward) than for the Australian (d
= +0.39, from zero to rightward) population.
The cultural differences in preferred object orientation have been attributed to reading habits, in that people prefer object pictures possessing the same directionality as their reading habits [4
]. However, there are several difficulties in this account of the preferred facing direction of single static objects. First, the Japanese language does not have the right-to-left reading direction in the same sense as the Hebrew and Arabic languages. The Japanese language can be read and written both horizontally and vertically. When writing vertically, the first line starts from the upper right of the page, and next lines follow to the left. These days, however, Japanese people are much more accustomed to the horizontal reading and writing direction, for instance, on computers and smartphones. The Japanese government has strongly encouraged the use of horizontal writing in public documents since 1952 [11
]. Therefore, the finding that Japanese individuals showed a strong leftward bias, which is the direct opposite to the rightward bias of English and French readers, cannot be explained by reading habits alone. If there were no other factors than reading habits on preference for the facing direction of static objects, Japanese people should have shown no or a weaker bias than English and French readers. In reality, Japanese people showed a larger leftward bias. The preference for left-facing object drawings was obtained for Taiwanese people, who also used horizontal and vertical reading systems [12
Besides the direction that the objects are facing, people often find that one view looks better or more “right” than its mirror-reversed view. This disposition is called “directionality” [4
]. Directionality appears in several domains, including movement direction and spatial compositions, and it has been proposed to be influenced not only by cultural factors, but also by biological factors. Biological factors include cerebral laterality (hemispheric asymmetry), which is associated with handedness. Because the spatial processing of pictures involves right hemisphere activation, attention tends to move to the left, and the processing of stimuli in the left hemifield, or the left side of a stimulus, is facilitated [14
The present study aimed to confirm the cultural differences in the preferred facing direction of centrally presented, static objects. In three online surveys, 100 pairs of symmetric objects and their mirror-images were presented. The participants were asked to choose the member that looked better or aesthetically pleasing. Survey 1 (a pilot survey) was conducted with Japanese university students (n
= 41) to replicate the leftward bias reported in Ishii et al. [10
]. Survey 2 (the main survey) recruited both American (n
= 492) and Japanese (n
= 505) participants and examined whether the opposite directional biases in English and Japanese readers were replicated with 10 times larger sample size than the original study [10
]. Survey 3 (a supplementary survey) asked a small group of Israeli participants (n
= 44) to answer the same questionnaire, here aiming to confirm the leftward bias in Hebrew readers [7
Moreover, using the large-sample data of the main survey (Survey 2), the effects of handedness and object type were also examined. Previous research on the relationship between handedness and preferred facing direction has produced inconsistent results. Nachson et al. [7
] reported that non-right-handed subjects showed more rightward preference than right-handed subjects for at least some types of stimuli, whereas De Agostini et al. [9
] reported the opposite. The preferred facing direction may also depend on the functionality of the object. For instance, an object with a handle may look better when it is displayed at the angle in which the handle is directed toward the observer’s dominant hand than at the opposite angle in which the handle is not easy to grasp. This tendency has been reported in a production task where participants were asked to draw a picture of objects freely; they tended to draw an object in the direction with its handle facing toward their dominant hand [17
]. Finally, a correlation between the preference indexes of Japanese and American participants was calculated at the object level. If their preferences were formed symmetrically (i.e., an object with a stronger leftward bias in one country had a stronger rightward bias in the other country), a negative correlation would be expected across objects. In contrast, if there was a base leftward or rightward preference in each country and relative differences among objects were determined by other factors, such as object functionality, a positive correlation would be expected.
In the current study, the preferred facing direction of three-dimensional symmetric objects was examined in three online surveys (total N
= 1082). The results showed that consistent with previous findings [7
], clear cultural differences can be found: The Japanese and Israeli participants preferred a leftward direction, whereas the American participants preferred a rightward direction. The effects of handedness and object type were also found. Left-handers showed more rightward bias than right-handers, irrespective of culture. Moreover, the angle of an object whose handle was easy to grasp with the dominant hand was preferred over the opposite angle. However, these effects were much weaker than the cultural differences.
The present study demonstrates that cultural factors are the major determinant of the perceptual preference of the facing direction of three-dimensional objects. However, the proportion of variance accounted for by culture was small and only 4.7% in Survey 2. The histograms (Figure 4
) show that a considerable number of individuals who had a strong leftward bias and who had a strong rightward bias exist in all groups, irrespective of culture. The large individual differences imply that the responses in this two-alternative forced-choice task are easily biased not only by culture, but also by other unspecified factors.
Although reading habits have been considered as a main cultural factor in directionality research [4
], reading directions do not seem to explain the cultural differences, at least in this perceptual evaluation task. As mentioned in the Introduction section, if one’s reading direction enhances the preference for the facing direction that matches this habit, unidirectional English readers would show a larger bias than multidirectional Japanese readers. This was not the case. The absolute effect size (i.e., deviation from zero) was much larger for the Japanese participants (d
s = −0.54 and −0.58 in Surveys 1 and 2, respectively) than for the American participants (d
= +0.19). This tendency is consistent with the results of Ishii et al. [10
s = −0.82 and +0.39 for Japanese and Australian individuals, respectively. The Israeli participants in Survey 3 also showed a strong leftward bias (d
= −0.75), although all of them declared that they also read English.
At the object level, the relative rankings of directional preference among the 100 objects were similar (i.e., positively correlated) between the Japanese and American participants. This means that an object with a strong leftward preference bias in one culture tends to have a relatively more leftward bias in the other culture, too. This finding supports the idea that each culture has a base bias, either leftward or rightward, and the preference for a certain object is determined by its attributes. The interaction effect of handle position and dominant hand also supports that object attributes affect directional preference, regardless of culture.
Taking all these findings into account, it may be reasonable to assume that a leftward bias, rather than the rightward bias observed in English and French readers, is the basic disposition in the facing direction preference of static objects that do not imply any movement. Although this idea directly comes from the results of the present and previous studies [10
] that Japanese and Israeli readers showed a stronger bias than English readers, it is supported by several lines of research in related fields. Three sources of evidence will be described below.
First, human faces tend to be depicted to be facing toward the left (i.e., showing the left cheek of the model) in drawings [20
], paintings [22
], and photographs [23
], with exceptions in self-portraits done in the mirror [26
], across cultures. Therefore, faces that turn to the left appear to be more familiar than faces in the opposite direction. Although this disposition has been replicated as a whole, which cheek is put forward is affected by various factors, including the model’s gender [22
], professions [28
], expressive intention [30
], painter’s handedness [13
], and reading habits [32
]. Because faces are processed by specialized neural mechanisms, the findings on face processing may not be applicable to more general visual processing [34
Second, the overall impression of a stimulus tends to be determined by the left side of the stimulus [16
]. In experiments using chimeric faces (e.g., one-half of the face is smiling and the other half is neutral), people are likely to choose the faces whose left side was smiling as a happier face [15
]. This leftward bias has been discussed in connection with cerebral laterality or hemispheric functional asymmetry. Research has shown that left oblique faces are recognized faster than right oblique faces [39
]. Given that attention is directed more easily to the left hemifield than to the right hemifield [14
], faces and objects would be processed more efficiently when their front side (presumably the main part) is turned to the left than right side [16
]. This leftward bias is shown to be more prominent in left-to-right (Hindi) readers than in right-to-left (Arabic or Urdu) readers or in illiterates [40
Third, people tend to assume that a light source is located at their upper left [41
]. An object is usually lit from its front to make it more clearly visible, as in the stimulus images used in the current study. Therefore, left-facing objects lit from the left should be more consistent with one’s assumption of the light source than right-facing objects lit from the right. In fact, Smith and Elias [45
] reported that people showed a preference for left-lit photographs with a clear light source originating from the upper left quadrant over its mirror-images and that the leftward bias was larger for left-to-right readers than for right-to-left readers.
These pieces of evidence suggest that even for left-to-right readers, a leftward bias would be a default disposition in the facing direction preference of static objects. However, contradictory evidence has also been obtained. Treiman and Allaith [46
] reported that children whose native language was Arabic (right-to-left readers) showed a rightward bias and preferred right-facing objects over left-facing objects, which is the direct opposite of other findings using the same task [7
]. Boiteau et al. [47
] examined the paintings of humans and animals on rocks in southern Africa created by a culture without a written language. They reported that regions with a rightward bias outnumbered regions with a leftward bias, although the absolute number of left-facing figures was nonsignificantly larger than that of right-facing figures. Although these findings also cast questions on the role of reading habits in the directional preference of objects, the precise mechanism of determining the preferred facing direction in each culture remains unclear. If stimuli are objects that imply a movement (e.g., arrows), the reading direction may affect the preference. However, when using stationary objects, the preference seems to be determined by other factors.
One possibility is that the definition of view goodness differs across cultures. For instance, Western culture often values symmetry as an aesthetic standard [48
], whereas the Japanese culture appreciates asymmetry, as typically shown in Japanese rock gardens [49
]. The preference for symmetry is also influenced by education and training on aesthetic appreciation [50
]. Levy [15
] hypothesized that people tend to compensate for the leftward attentional bias because of cerebral laterality by putting weight on the right side. When seeing a landscape, English, Japanese, and Taiwanese readers alike prefer a landscape in which the focus of interest is on the right side [8
]. However, when it comes to single static objects, Western people (particularly, non-experts in art [50
]) may tend to restore the balance between the left and right sides of the object by focusing more on the unattended right side, whereas Japanese and other Asian people may leave the imbalance as it is. In connection to this, Blanz et al. [1
] reported that German participants did not show any leftward or rightward bias when asked to take the best picture of an object, but showed a leftward bias when they were asked to express their natural mental image of the object. Therefore, the choice of the view angle may be affected by task requirements. If other types of questions, such as “looks more natural” or “fits more with your conceived image” rather than “looks better” are used, different results may be obtained. Japanese and Israeli individuals may choose the image that is more consistent with their intuitive image of the object when they are asked to choose the member that looks better.
Left-handers had more rightward bias than right-handers, irrespective of culture. Nachson et al. [7
] reported that non-right-handed subjects showed more rightward preference than right-handed subjects, irrespective of reading direction, at least for some types of stimuli (but see [9
] for the opposite trend in French readers). The present results are generally consistent with the literature that left-handers have a weaker tendency to process visuospatial information predominantly in the right hemisphere [15
]. The finding that the angle of an object whose handle was easily graspable was preferred over the opposite angle is consistent with the previous finding in the drawing task [17
]. Blanz et al. [1
] reported that German participants showed a preference for this functionally appropriate view in the mental image task, but not in the photography task. The correspondence between the dominant hand and the preferred angle of manipulative objects may be related to the concept of affordance [51
], which may induce the automatic activation of motor-related brain regions [52
]. Although the relationship between motor affordance and view goodness is not self-evident [53
], the angle of an object whose handle is graspable may be preferred because it is a more familiar view than the opposite angle.
There are some limitations to this research. First, the present study dealt with monochrome images of single objects. If more complex images with multiple items are used, the results may change. For example, Taiwanese people showed no bias when more realistic pictures were used [12
]. Second, it remains unclear if the observed bias can be generalized outside the two-alternative forced-choice task. For example, Niimi and Yokosawa [2
] did not find any directional bias when an object was presented alone, and its evaluation was made using a 7-point scale. The directional bias might be very weak, or a special phenomenon observed only in the two-alternative forced-choice task in which an image and its mirror-reversed version are directly compared [54
]. Third, the objects in this study were presented without any specific frame. Palmer et al. [55
] showed that the preferred facing direction of a single object depended on its position in the rectangular frame. That is, the object that faced into rather than out of the frame was preferred (i.e., inward bias) [56
]. It is worth examining whether this interactive effect of the facing direction and the position in the frame on aesthetic preference varies across cultures. Fourth, this study is purely descriptive and does not provide any process-based accounts. Further research is required to elucidate the underlying mechanisms. Among them, it is worth examining how consistent and stable one’s preference is across time and situation. Measuring the test-retest reliability of the preference index would be helpful to consider the basis of this preference. Another direction is to examine the effect of short-term experience on the preferred facing direction. If familiarity is a major determinant of view goodness [3
], repeated exposure to objects with a certain facing direction would increase the likelihood of choosing the angle that is consistent with the preceding views [57
In conclusion, the current study confirms that the preferred facing direction of a symmetrical stationary object displayed obliquely is influenced by cultural factors. The answer to the opening question, “Which side of a three-dimensional object looks better?” may be “it depends on the culture.” Handedness and the functionality of the object are also shown to affect the preference, but only slightly. In addition, the available evidence suggests that a left-facing view could be a standard angle of static objects, although it would not be called a “good” view.